Model Cost Check
Next changesGPT-5.6 Sol promo can endGemini 4 Argon+100%Gemini 3.8 and 3.6 Flash+100%Prices checked

Home / Compare / DeepSeek vs ChatGPT

DeepSeek vs ChatGPT

DeepSeek’s app is free, with no paid plan at all, and its model’s weights are public. ChatGPT has a free app too, plus paid plans from $8 to $500 a month, and is used vastly more. Through the API, DeepSeek is no longer the cheapest: OpenAI’s GPT-6 Luna now costs less per token, except for cached input. DeepSeek also stores your data in China.

Checked October 18, 2026 against DeepSeek’s pricing page, its App Store listing and ChatGPT’s pricing page

DeepSeek and ChatGPT side by side
DeepSeekChatGPT
App priceFree; no paid plansFree, or Go $8, Plus $20 and Pro from $100 a month
Newest modelDeepSeek V4.1 Flash (updated October 17, 2026)GPT-6 Astra
Open weightsYes, published on Hugging FaceNo
ImagesReads images (V4.1 Flash); no image generation listedImage generation and editing in the free app
Where your data is storedIn the People’s Republic of China, its privacy policy saysSee OpenAI’s privacy policy
Share of chatbot visits to websites0.02%80.48%

ChatGPT prices from OpenAI’s pricing page. Share of visits: Statcounter, September 2026.

The apps: free against paid plans

DeepSeek’s assistant app is free and has no in-app purchases; you can also chat on its website. ChatGPT’s app is free too, with image generation, voice and photo upload; its paid plans are Go at $8, Plus at $20 and Pro from $100 a month.

DeepSeek also publishes its models openly: the weights of DeepSeek V4.1 Flash are on Hugging Face, so anyone with the hardware can run it themselves. OpenAI’s GPT-6 models are available only through OpenAI and its partners.

DeepSeek and ChatGPT plans by price

  1. FreeChatGPTFree
  2. FreeDeepSeekFree
  3. GoChatGPT$8 a month
  4. PlusChatGPT$20 a month
  5. ProChatGPTfrom $100 a month

US prices a month before tax. DeepSeek’s app has no paid plan; ChatGPT Pro is the lowest of several tiers.

DeepSeek vs ChatGPT models through the API

DeepSeek charges by time of day: its listed prices apply at peak hours (01:00–04:00 and 06:00–10:00 UTC, Monday to Friday, except Chinese public holidays) and everything else costs half. Even off-peak, DeepSeek Flash costs more per token than OpenAI’s GPT-6 Luna for fresh input and output; only cached input, the text a model has already read, is cheaper at DeepSeek.

Per million tokens
ModelInputCached inputOutput
DeepSeek Flash, peak$0.30$0.006$1.20
DeepSeek Flash, off-peak$0.15$0.003$0.60
DeepSeek V4 Pro, peak$1.32$0.044$3.96
GPT-6 Luna$0.10$0.01$0.50
GPT-6.1 Sol$2.00$0.10$10.00
GPT-6 Astra$10.00$1.00$50.00

Prices side by side

DeepSeek FlashDeepSeek
GPT-6 LunaOpenAI
DeepSeek Flash: $0.30InputDeepSeek Flash costs 3×GPT-6 Luna: $0.10
DeepSeek Flash: $0.006Cached inputGPT-6 Luna costs 1.7×GPT-6 Luna: $0.01
DeepSeek Flash: $1.20OutputDeepSeek Flash costs 2.4×GPT-6 Luna: $0.50
DeepSeek Flash: $27.18A month, support-chat exampleDeepSeek Flash costs 2.5×GPT-6 Luna: $10.80

US dollars per 1 million tokens, standard tier, from each company’s pricing page. The month: 30,000 requests of 2,000 tokens in (half read from cache) and 500 out. Set your own numbers in the calculator below.

What it costs in a month

$27.18DeepSeek Flash at peak rates, 30,000 chats of 2,000 tokens in (half cached) and 500 out
$13.59DeepSeek Flash, all off-peak
$10.80GPT-6 Luna, same use
$1,080GPT-6 Astra, same use

Prices aren’t the whole story: OpenAI positions Luna for cost-sensitive, high-volume work and Astra for its hardest tasks, so compare results on your own tasks.

Start from an example
Count in

A token is a piece of a word. For English, 100 tokens is about 60 to 80 words (Google). In: your prompt, instructions and any pasted text. Out: the answer, including any thinking the model does before it answers. Cache: the part of each prompt that repeats, such as fixed instructions, billed at the lower cached rate.

DeepSeek vs GPT for your use

Sort
Choose models
  1. Turn on JavaScript to see the costs. The price tables below work without it.

Privacy: where your chats go

DeepSeek’s privacy policy says it collects, processes and stores your personal data in the People’s Republic of China, and that it may be stored on servers outside the country where you live. If that matters for your work, the open weights offer a way round it: run the model on your own hardware or with a host you choose.

What people say

The DeepSeek V4.1 Flash launch drew more than 570 comments on Hacker News. Developers praised the detailed technical report and the very low price of cached input, and one called it a strong, cheap backup for when a paid subscription runs out. On the other side, users said the app sometimes answers English questions in Chinese, and one pointed out that DeepSeek’s own benchmark table puts it ahead of OpenAI’s older GPT-5.6 Sol on some tests and behind on others.

DeepSeek’s share of the website visits sent by AI chatbots was 0.02% in September 2026, against 80.48% for ChatGPT (Statcounter).

Questions

Is DeepSeek free to use?

Yes. DeepSeek’s app is free and has no paid plan. Developers pay per token through the API, from $0.15 per million input tokens off-peak.

Is DeepSeek better than ChatGPT?

It is cheaper to use and its weights are open, but ChatGPT makes images, has voice mode and is far more widely used. For developers, DeepSeek Flash’s cached input is the cheapest of any model here, while GPT-6 Luna is cheaper for fresh input and output.

Is DeepSeek safe to use?

DeepSeek stores your data in China, its privacy policy says. Don’t share anything you wouldn’t want held there; for sensitive work, run the open model yourself or use a host you trust.

Is DeepSeek cheaper than ChatGPT through the API?

Not always. In our monthly example, DeepSeek Flash costs $27.18 at peak rates and $13.59 off-peak, GPT-6 Luna $10.80 and GPT-6 Astra $1,080.

Sources